Chapter 3 • Telecommunications and Networking 73BusPC PCPCPC PC PCPCHubFIGURE 3.6 Shared Ethernet Topology: Logical Bus,
Physical StarA LAN differs from a computer telecommunica-
tions network in that a LAN contains a number of intelli-
gent devices (usually microcomputers) capable of data
processing rather than being built around a central com-
puter that controls all processing. In other words, a LAN
is based on a peer-to-peer relationship, rather than a
master–subordinate relationship.
There are five types of LANS in use today—three
types of wired LANs and two types of wireless LANs—for
which standards have been developed by the Institute for
Electrical and Electronic Engineers (IEEE) and subse-
quently adopted by both national and international
standards organizations. These five LAN standards are
officially designated as IEEE 802.3 (contention bus
design); IEEE 802.4 (token bus design); IEEE 802.5
(token ring design); IEEE 802.11, including 802.11a,
802.11b, 802.11g, and 802.11n (Wi-Fi wireless design);
and IEEE 802.16, including 802.16d and 802.16e
(WiMAX wireless design).
Wired Local Area Networks. Thecontention bus
design was originally developed by Xerox and subsequent-
ly adopted by Digital Equipment Corporation (now part of
Hewlett-Packard) and several other vendors. This design is
usually referred to as Ethernet, named after the original
Xerox version of the design. The contention bus is
obviously a bus topology (see Figure 3.4), usually
implemented using coaxial cable or twisted pair wiring.
Communication on an Ethernet LAN is usually half-
duplex—that is, communication in both directions is
possible, but not simultaneously. The interesting feature of
this design is its contention aspect—all devices must
contend for the use of the cable.
With Ethernet, devices listen to the cable to pick off
communications intended for the particular device and
determine if the cable is busy. If the cable is idle, any
device may transmit a message. Most of the time this
works fine, but what happens if two devices start to
transmit at the same time? A collision occurs, and the
messages become garbled. The devices must recognize
that this collision has occurred, stop transmitting, wait a
random period of time, and try again. This method of
operation is called a CSMA/CD Protocol, an abbrevia-
tion for Carrier Sense Multiple Access with Collision
Detection. In theory, collisions might continue to occur,
and thus there is no upper bound on the time a device
might wait to send a message. In practice, a contention bus
design is simple to implement and works very well as long
as traffic on the network is light or moderate (and, thus,
there are few collisions).
The original Ethernet design, now called shared
Ethernet, employs a contention busas its logical topology,
but it is usually implemented as a physical stararrange-
ment (see Figure 3.6). The usual way of creating a shared
Ethernet LAN is to plug the cables from all the devices on
the LAN into a hub, which is a junction box containing
some number of ports (e.g., 12) into which cables can be
plugged. Embedded inside the hub is a linear bus connecting
all the ports. Thus, shared Ethernet operates as a logical
bus but a physical star.
Switched Ethernetis a newer variation of Ethernet
providing better performance at a higher price. The design
is similar to shared Ethernet, but a switch is substituted for
the hub and the LAN operates as a logical star as well as a
physical star. The switch is smarter than a hub—rather
than passing all communications through to all devices on
the LAN, which is what a hub does, the switch establishes
separate point-to-point circuits to each device and then
forwards communications only to the appropriate device.
This switched approach dramatically improves LAN
performance because each device has its own dedicated
circuit, rather than sharing a single circuit with all devices
on the network. Of course, a switch is more expensive than
a simple hub.
Thetoken busdesign employs a bus topology with
coaxial cable or twisted pair wiring, but it does not rely on
contention. Instead, a single token (a special communication
or message) is passed around the bus to all devices in a spec-
ified order, and a device can only transmit when it has the
token. Therefore, a microcomputer must wait until it receives
the token before transmitting a message; when the message
is sent, the device sends the token on to the next device. After
some deterministic period of time based on messages sent by
other devices, the device will receive the token again.